Re: [PATCH 09/18] tests/qtest/migration: Add a test for HMP

[Peter Xu <peterx@redhat.com>]

On Thu, Sep 03, 2026 at 04:38:43PM -0400, Peter Xu wrote:
> On Wed, Sep 02, 2026 at 07:15:37PM -0300, Fabiano Rosas wrote:
> > The following patches will change how parameters are set and shown in
> > HMP, including readline completion so add two minimal tests.
> > 
> > - set/info migrate_parameters
> > 
> > The test uses qtest facilities to issue HMP migrate_set_parameters for
> > each of the existing migration parameters and queries them back with
> > the info command. A list is kept with the expected strings. A
> > substring match function inspired by glib's g_str_match_string is
> > implemented for this test so the test can produce a decent error
> > output instead of just assert failure. E.g:
> > 
> >  # HMP output mismatch for entry at line 55:
> >  # expected vs. found:
> >  #
> >  # max-bandwidth: 10356305952768 bytes/hour
> >  # ---
> >  # max-bandwidth: 10356305952768 bytes/second
> > 
> > (note that line 55 above is the source line where the test case for
> > max-bandwith is)
> > 
> > - readline completion
> > 
> > The test puts the monitor on a chardev via socket and bypasses qtest
> > facilities because it needs to emit raw codes to readline. It
> > therefore requires a couple of new helpers to read/write to the
> > monitor socket.
> > 
> > Usage:
> > QTEST_QEMU_BINARY=./qemu-system-x86_64 \
> > ./tests/qtest/migration-test --full -p /x86_64/migration/hmp
> > 
> > Signed-off-by: Fabiano Rosas <farosas@suse.de>
> 
> Acked-by: Peter Xu <peterx@redhat.com>

Ah I forgot to say: Logically this test should not be in migration-test,
simply because it uses 1 VM not 2.. hence MigrateCommon doesn't make sense
to be passed into the two HMP tests.  But can be done later too.

Also feel free to also keep my A-b if you lightly update it, e.g. add
set-parameter test entries, duplicate g_test_message()s, etc.

-- 
Peter Xu